Small antibodies, big libraries

If you work in the antibody discovery space, you’ve probably considered using synthetic libraries at some point. You might have heard about the speed at which they can generate hits, but you’ve probably also heard some horror stories about the quality of the antibodies produced in this way.  

As VHH experts, we’re big proponents of using synthetic libraries for antibody discovery. But what exactly are the benefits? Do they even produce good molecules? And what do you gain by having access to the largest and most diverse VHH libraries on the market?  

The truth about synthetic libraries

One of the most common misunderstandings around synthetic libraries is that they produce poor quality, low affinity antibodies. Though this may be true for some of the synthetic libraries you’ve encountered in the past, it is simply not the case for all of them.   

The reason this misconception lives on today is because various service providers have previously overhyped the claims about the capabilities of their synthetic libraries, and they’ve simply not lived up to the promise. These providers have ended up disappointing their partners, who poured a lot of time, money, and effort into discovering antibodies, only to end up with poor quality products that can’t be used in a clinical setting.  

In reality, with a deep understanding of what a VHH is and how it works – because it’s quite different to a standard antibody – it is possible to achieve high quality, high affinity single-domain antibodies (sdAb) with synthetic libraries.  

In fact, the bestselling antibody drug of all time, HUMIRA®, was engineered using a synthetic library. This goes to show that by asking the right questions of the right synthetic library, you can produce fantastic results.  

Synthetic VHH libraries versus traditional immunisation

One major advantage of synthetic libraries is that they are much faster than traditional immunisation. Antibody discovery via immunisation can take up to two years to recover a usable antibody, whereas a synthetic approach can take as little as 7 months.  

A good immune library requires a lot of target antigen, and a lot of waiting, boosting, and titre checking. Even after all this, your high-potency llama antibody may have a very high chance of raising an immune response in a human host.  

You can then end up ruining all your hard work in the process of humanising your llama antibody – the molecule can start playing up, falling apart and sticking to things you don’t want it to. So there’s a lot of inherent risk with immunisation, and it takes a lot longer.   

You might also run into issues around your freedom to operate or patentability with traditional immunisation approaches. It turns out that if you inject two different llamas with the same target protein, they actually stand a very good chance of producing very similar antibody sequences in response. This is less of an issue for novel targets, but many well-established drug targets have already been discovered in llamas. This means careful review of your sequences is needed at an early stage, to avoid making a non-novel VHH antibody you can never patent and commercialise.  

With synthetic libraries – especially gigantic ones like ours – the antibodies produced are going to be completely different as they haven’t gone through the same biological machinery. You are able to have thousands more shots on goal, different selective pressures, and therefore more chance of finding something with a truly novel sequence.  

The impact of library size on antibody affinity

The reality is, anyone can build a synthetic library – but it’s difficult to build a good one. The key to creating a high quality VHH library lies in understanding the inherent differences between VHHs and VH domains from standard antibodies.   

VHHs, often called nanobodies, represent the entire antigen binding region of a heavy-chain-only antibody. Compared to a standard antibody, which has six complementarity-determining regions (CDRs), a VHH has just three. This means it has to work twice as hard to achieve the high affinities seen with standard antibodies.  

Now, a llama or camel has developed its machinery over millions of years of evolution to land on the perfect system for creating high affinity VHH antibodies. But recreating this synthetically has proved tricky for many. So how can we overcome this challenge and produce high affinity, single domain antibodies?   

When it comes to synthetic DNA libraries, size matters. In fact, a recent review concluded that the biggest factor in obtaining high affinity antibodies is CDR sequence diversity. In other words, the bigger the library, the more chance of producing really high affinity binders, and the better your final product.  

By using a library size of up to around 10^12, you can start to achieve really high affinities because you’ve got a lot more choice. If you have more choice, you can essentially force evolutionary conditions upon the antibodies by limiting the number of target molecules and getting them to compete with each other. All of a sudden, there is a crowd of these antibodies fighting for position on the target, and you can select the best ones.    

What makes Isogenica’s libraries so special?

At Isogenica, our libraries are around 100-1000 times bigger than most other synthetic VHH libraries (Figure 1).

In fact, they are the largest and most diverse synthetic VHH libraries on the market and we 

This has huge implications in terms of affinity, developability, and freedom to operate. By simply providing more shots on goal, our libraries are able to achieve the same high affinities you would expect from animal immunisation campaigns, but with the speed of animal-free antibody discovery (Figure 2).

At Isogenica, we specialise in VHHs, meaning all our libraries are based on this format. We have 3 different libraries launched so far; LlamdA®, huLlamdA®, and VHHantage® (Figure 3). We have engineered each library to specifically suit the production of therapeutics  and are always iterating to make further improvements based on what we learn. 

Figure 2: Typical ELISA EC50s of VHH antibodies discovered in Isogenica’s antibody discovery campaigns. All campaigns were against cell-surface displayed antigens and also passed cell binding assessments. With the exception of Target 5 (an extremely small antigen), all campaigns generated single digit nM binders, with 3 out of 7 generating sub-nM binders.

Each library has been engineered to have reduced CMC liabilities such as free cysteines and glycosylation or deamidation sites, reducing the downstream engineering work required to turn immune-derived VHHs into potential drugs. What’s more, our huLlamdA® and VHHantage® libraries are pre-humanised, further cutting down on engineering time. By using all three in parallel, it allows us to expand the repertoire of possible binders and further increase our chances of finding the perfect match.  

Though we’ve experimented with other cassettes over the years, such as human IgG-derived CDRs, we’ve learned that part of what enables VHHs to concentrate all their binding ability into just one domain is their unique CDR compositions. That’s why all our libraries are based on natural distributions, instead of cutting and pasting existing fragments.

A powerful combination: synthetic libraries and CIS display  

Having a big library is great, but it’s only an advantage if you are able to interrogate that huge diversity. Thanks to our proprietary CIS display technology, we can interrogate up to 1000 times more VHH antibody sequences than standard display technologies.  

CIS display is a similar concept to phage display – it’s a way of labelling a protein with its own genetic information. However, unlike phage, yeast, or mammalian display, it does not involve a biological system, meaning library sizes are less restricted and huge libraries can be mined effectively.  

With traditional phage display, libraries are limited by the efficiency of getting DNA across the membrane. Going up an order of magnitude in terms of library size is essentially 10 times more work, and 10 times the cost. The beauty of CIS display is that the DNA is not part of a virus (bacteriophage) genome, it is just a PCR product that you drop into a mixture to create protein-DNA complexes (Figure 4).  

Figure 4: How CIS Display works. A DNA construct is created where the library is encoded upstream of the repA gene, followed by a CIS-ori sequence. During in vitro transcription and translation, the RepA protein binds to the CIS-ori genetic element of its own transcript, thus linking the translated VHH from the library with its own genetic information. These fusions can then be used directly in panning. Critically, rounds of selection can be done in the absence of a biological system, meaning library size and diversity are less restricted.

Looking for a partner in small antibody discovery?  

At Isogenica, we have decades of experience in discovering antibodies and peptides for our partners. Our knowledge and experience of VHHs has grown to the point where, as The VHH Company™, we can fully realise the potential of our huge and diverse synthetic VHH libraries, helpings us deliver better antibodies faster, with higher affinity and broader IP coverage. If this sounds like what you’re looking for, get in touch with our team today.   

Still not sure whether VHHs are for you? Explore our blog and technical resources where we share our expertise on the discovery and engineering of highly versatile single domain VHH antibodies. Or read more about what we’ve done for our current and past partners.